Phase II Final Report DVMCS P

DOWNHOLE VIBRATION MONITORING & CONTROL SYSTEM FINAL REPORT – PHASE II

Starting date: June 1, 2004 Ending Date: January 31, 2006 Principal Author: Martin E. Cobern Date Issued: April 30, 2006 DOE Award Number: DE-FC26-02NT41664 Submitting Organization APS Technology, Inc. 800 Corporate Row Cromwell, CT 06416

Phase II Final Report DVMCS p. 1 DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not in- fringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Govern- ment or any agency thereof.

ABSTRACT The objective of this program is to develop a system to both monitor the vibration of a bottomhole assembly, and to adjust the properties of an active damper in response to these measured vibrations. Phase I of this program, which entailed modeling and design of the necessary subsystems and design, manufacture and test of a full laboratory prototype, was completed on May 31, 2004. The principal objectives of Phase II were: more extensive laboratory testing, including the evaluation of different feedback algorithms for control of the damper; design and manufacture of a field prototype system; and, testing of the field prototype in drilling laboratories or test wells. The specific tasks were modified in November, 2005 and these tasks are used as the basis of organization for this report. The laboratory testing at TerraTek Laboratories was completed in January, 2006. These tests demonstrated that the DVMCS can maintain more consistent weight- on-bit, decrease vibration and increase the rate of penetration. With the exception of Task 10 (Ordering all long lead items for field prototypes), all of the tasks outlined have been completed and form the basis for the field testing and commercialization in Phase III, which is scheduled to end in January, 2007.

Phase II Final Report DVMCS p. 2 Table of Contents Executive Summary...... 4 Task 1: Develop final prototype development & testing plan...... 5 Task 2: Complete detailed design of prototype system using CAD...... 5 Redesign of laboratory prototype ...... 5 Design of feedback system ...... 5 Intermediate prototype design ...... 5 Task 3: Build mockups for laboratory testing...... 6 Task 4: Test mockups in laboratory and analyze performance ...... 6 Preliminary Tests & Analysis...... 6 Implications for Redesign ...... 7 Further testing with demagnetization...... 7 Task 5: Redesign mockups based on test results ...... 9 Task 6: Build revised mockup ...... 9 Task 7: Retest mockups in laboratory and analyze performance ...... 10 Task 8: Refit laboratory mockup for use in drilling laboratory...... 10 Task 9: Test in drilling laboratory...... 13 Task 10: Complete procurement of long-lead items...... 14 Task 11: Update economic, market and environmental analyses...... 15 Economic & Market Analysis...... 15 Environmental analysis ...... 15 Task 12: Update financing plan...... 15 Task 13: Submit Phase II final report ...... 16 Other ...... 18 Units ...... 18 References...... 18 Appendix A: Minimum WOB Variation...... 19 Appendix B: Hardening Algorithm ...... 24 Appendix C: Magnetic Testing ...... 30 Appendix D: MR Fluid Viscosity Testing ...... 37 Appendix E: Testing of DVMCS Prototype with Outer Coils ...... 47 Scope...... 48 Conclusions...... 48 Discussion...... 49 Outer Coil Results...... 50 Original Inner Coil Results...... 53 Appendix F: Analysis of TerraTek Results...... 54 Scope...... 55 Conclusions...... 55 Recommendations ...... 56 Discussion...... 56 Appendix G: Final Design of Field Prototype...... 69 Appendix H: Excerpt from Spears Market Study...... 71

Phase II Final Report DVMCS p. 3 Executive Summary The objective of this program is to develop a system to both monitor the vibration of a bottomhole assembly, and to adjust the properties of an active damper in response to these measured vibrations. Phase I of this program, which entailed modeling and design of the necessary subsystems and design, manufacture and test of a full laboratory prototype, was completed on May 31, 2004. The principal objectives of Phase II were: more extensive laboratory testing, including the evaluation of different feedback algorithms for control of the damper; design and manufacture of a field prototype system; and, testing of the field prototype in drilling laboratories or test wells. The specific tasks were modified in November, 2005 and these tasks are used as the basis of organization for this report. The laboratory testing at TerraTek Laboratories was completed in January, 2006. These tests demonstrated that the DVMCS can maintain more consistent weight- on-bit, decrease vibration and increase the rate of penetration. With the exception of Task 10 (Ordering all long lead items for field prototypes), all of the tasks outlined have been completed and form the basis for the field testing and commercialization in Phase III, which is scheduled to end in January, 2007.

Phase II Final Report DVMCS p. 4 Task 1: Develop final prototype development & testing plan This task was completed early in this Phase, and then modified slightly in November, 2005. The tasks below represent this revised plan.

Task 2: Complete detailed design of prototype system using CAD

Redesign of laboratory prototype The redesigned valve now has the coils located in the non-reciprocating portion of the tool. This lends itself to a more reliable electrical connection and better protection of the coils from the abrasive MR fluid.

Design of feedback system The laboratory results of Phase I were analyzed to determine which particular feedback algorithms, and which input data, were likely to result in the most efficient control of the DVMCS. Two algorithms were identified for further study using the laboratory prototype: ? Minimum WOB variation. In this algorithm, the relative motion of the two halves of the DVMCS is used as a proxy for the approximate WOB. The algorithm minimizes the change in DVMCS motion, thereby keeping WOB constant. A memo describing this algorithm and its implementation is attached as Appendix A. ? Hardening algorithm. The first algorithm may have some difficulties near the ex- tremes of motion. Under certain conditions, it might cause the DVMCS to ‘lock up’ and effectively remove all damping from the system. To remedy this possible problem, a ‘hardening algorithm’ was developed, which uses a quadratic factor, also based on the relative motion of the DVMCS. This approach is described in Appendix B.

Intermediate prototype design The design of the field prototype evolved considerably over the course of this phase. Among the areas added or changed in the design are: ? Addition of a battery-powered, self-contained unit to record accelerations at the bit. (This is for evaluation purposes, and will likely not be a part of the commercial tool.) ? Addition of a battery to the DVMCS to preserve the absolute position when the tool is powered down. ? Elimination of the WOB sensor, as we will use the absolute deflection of the DVMCS as a proxy for this measurement. ? Development of a connector to transfer power and data between the turbine- alternator unit and the DVMCS sub.

Phase II Final Report DVMCS p. 5 According to the revised Statement of Work, the laboratory prototype design was modified to allow the prototype to be run in the drilling laboratory at TerraTek. Among the changes were: ? Replace the instrumented “bit” element with the appropriate bit box. ? Install the battery-operated vibration monitoring sub ? Install the internal motion controller input. ? Manufacture new upper sub to interface with the TerraTek commutator.

Task 3: Build mockups for laboratory testing The revised laboratory mockup was built with the modifications described above. it was tested on our test bench, using the sweep algorithm. During testing, it was noted that the motor and gear train of the test bench could not provide sufficient power at higher frequencies without overheating or tripping the controller circuit breaker. Testing was therefore limited to 1.5 Hz. The initial results indicated a deviation from our earlier static testing and the model predictions. Given these two problems, testing of the feedback algorithms was postponed until the causes were better understood. A new gearbox/motor combination was designed and installed in the test bench. A viscometer was procured, and was modified to include magnetic coils, to study the properties of our ‘home-made’ MR fluid vs. the Lord commercial fluid. Testing showed no differences in the fluids. (See below.)

Task 4: Test mockups in laboratory and analyze performance

Preliminary Tests & Analysis A preliminary analysis of the dynamic test data showed variations in the dynamic stiffness of the damper varied over ranges from 11% (at 5,000 lbs. WOB) to 80% (at 10,000 lbs. WOB). These are significantly below the ranges predicted by the modeling and earlier static tests. Several explanations were posited for these findings: ? It was possible that the passive parts of the system – the oil-filled Belleville springs, the compensation system, and general friction – themselves provided significant dynamic stiffness, which reduces the relative effect of the DVMCS. To study this effect, the MR fluid was drained from the valve the system response using only the passive components. This test showed that viscous damping in the spring stack had a negligible effect on the overall damping, ruling this out at as cause of the lack of range. The tests did show, however, that the spring rate was slightly higher than assumed. The modeling was modified to reflect this fact. ? The difference in the results could be attributed to differences in the magnetic permeability of the alloys used in the damper construction, which result in lower magnetic fields than predicted. To study these results, several parallel ap- proaches were used. o The magnetic fields in the gaps was measured directly.

Phase II Final Report DVMCS p. 6 o Additional modeling was performed and alternate coil designs studied. o The permeability of the alloys was measured directly. Analysis of the test data showed that the field in the damper gaps was only 60- 90% of that required for saturation of the MR properties of the fluid. The increase in viscosity of the fluid was also less than was predicted by the Lord literature. The results of this analysis are in Appendix C, which also compares the results with the magnetic modeling of the damper. The residual magnetic fields around the components were measured to be quite high, even in low coercivity ‘nonmagnetic’ materials. The low coercivity means, however, that a small reverse field can remove this residual magnetism. These tests and analyses are summarized in Appendix C, below. ? We used a ‘home made’ MR fluid for our testing, since the commercial fluid dis- played unsatisfactory settling properties. It was possible that this fluid does not display the range of viscosities under the influence of the magnetic field that the commercial fluid does. To evaluate this possibility, we purchased a commercial viscometer, and modified it to allow the measurement to be made while applying a magnetic field to the fluid. Also, the properties of both the Lord and ‘home made’ MR fluids were analyzed for their viscosity properties. The results of these tests are included in Appendix D, below In short, the tests showed that the home made MR fluid was comparable in properties to the commercial one, and that it allowed us to adjust the magnetic properties of the fluid by varying the ratio of iron to base oil. Based on these results, the laboratory prototype external control circuitry was modified to include an demagnetization algorithm, and the earlier testing was repeated. The results were markedly improved and are presented below.

Implications for Redesign The above tests led to several obvious conclusions: ? The damping of the inert components is not a significant factor in the overall damping coefficient of the DVMCS. ? The MR fluid used does not change its properties in a manner that would explain the lack of dynamic range observed in the first tests. ? The most significant cause of the results observed earlier was the residual magnetization of the components of the damper.

Further testing with demagnetization Based on the above, a demagnetization algorithm was built into the laboratory control apparatus and our earlier dynamic testing was repeated. The demagnetization process was extremely effective as illustrated by Figure 1, below.

Phase II Final Report DVMCS p. 7 AVD Demagnetization

16 14 12 10 8 6 Before Magnetization

Thousands 4 2 After De-Mag

Dynamic Stiffness (lb/in) 0 0 0.5 1 1.5 2 2.5 Freq (Hz)

Figure 1: Effect of demagnetization on performance of MR damper

With the demagnetization circuit in place, a dynamic range of 7:1 was achieved at lower frequencies, and 10:1 at higher frequencies, as seen in Figure 2, below. This represents a significant improvement over the earlier results.

AVD Dynamic Stiffness 5000 lbs WOB 140

120

100

60 0.5 Hz Thousands 40 1.0 Hz 1.5 Hz

Dynamic Stiffness (lb/in) 20 2.0 Hz 0 0 1 2 3 4 Current - amps

Figure 2: Performance of MR damper with demagnetization circuit operating

Phase II Final Report DVMCS p. 8 Task 5: Redesign mockups based on test results Based on the above observations, several changes were made in the design. ? The coil circuit was modified to demagnetize the valve components when the field is to be reduced. ? In order to allow a wider choice of materials for the valve components, the valve was moved above the Belleville spring. This decreases the amount of load the valve must bear and allowed the alloys to be chosen on the basis of their magnetic properties and resistance to damage by the mud, rather than particularly on their strength. ? The damper design was ‘inverted’, putting the coils in the outer housing and leav- ing the mandrel a constant diameter. Calculations show that this geometry will generate much higher fields. It also greatly simplifies assembly, since the mandrel, which must be carefully slid into the housing, will have an even surface. A sketch of the new design is shown below in Figure 3 . The coils are the orange features in the sketch.

Figure 3 Schematic of redesigned MR damper section

Task 6: Build revised mockup The revised mockup was constructed with the changes described above.

Phase II Final Report DVMCS p. 9 Task 7: Retest mockups in laboratory and analyze performance The reworked prototype, with external coils, was tested on our laboratory test bench, described in the Phase I Final Report1. (A picture of the test bench is included in Refer- ence 3.) The test data were analyzed by our analytical specialist, Mark Wassell, and his conclusions are included in Appendix E. The dynamic range of the revised damper circuit was less than that of the original. This was attributed to two factors: ? The reworked mandrel was slightly smaller than the original, which resulted in a larger gap. This, in turn, reduced the ‘power off’ damping coefficient and also reduced the applied field for a given current. ? The reworked mandrel still had some of the internal structure of the original coil winding slots. While these were filled in, the interfaces might interfere with the magnetic flux lines, decreasing the efficiency of the magnetic circuit. Despite these results, it was decided to go ahead with the testing using the current design. Since the ‘power off’ coefficient was quite low, the gap was reduced to increase the dynamic range. Time considerations prevented the manufacture of a new mandrel and the testing proceeded with the existing one.

Task 8: Refit laboratory mockup for use in drilling laboratory The new and revised components of the prototype were manufactured and assembled with the changes described above. The test procedure and matrix were developed, and four test formations were designed and built. These each consist of a slab of hard granite mounted at an angle of 10º within a larger hard concrete block. (See Figure 4- Figure 8.) The contrast in hardness at the inclined interfaces was designed to induce significant vibration in the drilling, which will serve as a test of the efficiency of the damper and its feedback algorithms. [Note that the holes shown in Figure 4 are schematic only, and do not represent the planned drilling pattern.]

Phase II Final Report DVMCS p. 10

Figure 4: Drawing of test blocks for drilling lab testing

Figure 5: Granite blocks and concrete molds ready for assembly

Phase II Final Report DVMCS p. 11

Figure 6: Positioning granite block at 10º angle before concrete pour

Figure 7: Pouring concrete for first block

Phase II Final Report DVMCS p. 12

Figure 8: Finished blocks begin setting process

Task 9: Test in drilling laboratory The tests were run from January 23-27, 2006. A summary of the results is included as Attachment F. The conclusions reached from analysis of the data include: 1. The vibration levels measured through the tests were fairly benign. These were in the 5 – 25 g level. 2. Significant bit wear occurred during the tests. For a conventional collar, the ROP at the end of the tests was 60% of the ROP at the start of the tests. To account for this the ROP rates were corrected based on a linear degradation of the bit for each hole drilled. Other drilling results, such as WOB, TOB acceleration, have not been corrected for bit wear. 3. Even with the benign drilling conditions the DVMCS showed its ability to improve ROP. While drilling through concrete, the DVMCS improved ROP by 10 – 15%. For granite the improvement was up to 11%. While drilling through granite at 120rpm with 15,000 WOB the ROP actually dropped by 2%; however, lighter damping might improve this situation. The lighter damping case was not run. 4. The optimum drilling condition occurs when the DVMCS internal travel is ~0.17”. If the drilling becomes too smooth, the ROP actually drops. There appears to be an optimum travel and WOB fluctuation that produces the best ROP. We believe that some vibration may improve cutting efficient, but this has not been proven.

Phase II Final Report DVMCS p. 13 5. The DVMCS tool significantly decreased the WOB fluctuations compared to the conventional drill collar. The DVMCS reduced the WOB fluctuation by 60%. This should be even more beneficial under high vibration conditions. 6. The DVMCS displacement was 0.25” for the power off state and 0.13” at full power. The dynamic stiffness varied from 4200 lb/in to 17,800 lb/in. These ranges should significantly improve for the commercial tool. Upon disassembly of the DVMCS tool, it was found that the brass bobbins for the damper coils had become slightly magnetic. This would have reduce the performance of the magnetic coils. 7. The tool was disassembled after the completion of the tests to inspect the seals and bearings. Both the bearings and the seals were still in good working order

Task 10: Complete procurement of long-lead items The testing described in Appendix I resulted in some considerable changes in the design of the prototype. These include the following: ? The internal position monitor (LVDT) did not produce sufficient signal within the collar to control the feedback loop, and it was necessary to rely upon an external sensor during the TerraTek testing. This has necessitated a complete redesign of the LVDT, which is currently underway. Tests with a ¼-scale model were very promising and a full-scale laboratory prototype is currently being assembled. ? Assembly of the laboratory prototype, and its conversion to the drilling lab prototype, demonstrated that the current design was overly complicated and extremely difficult to assemble, and not commercially viable. ? Furthermore, potential commercial partners indicated a strong preference for a system which can be integrated into their existing shock subs. A significant part of the DVMCS (bearings, Belleville springs, etc.) is very similar to a shock sub, with the key addition being the active feedback and control system. Based on these considerations, the use of our existing prototype in the field was not considered practical and began a redesign. The result is a much improved tool. In particular: ? The hydraulic compensation system was modified so that it uses a single reservoir at the top of the tool. By eliminating the lower reservoir, the bottom end of the tool becomes essentially identical to a standard shock sub. This permits its integration into existing products and will greatly reduce manufacturing costs. ? A reconfiguration of the tool permits greater flexibility in assembly and maintenance. In particular, it will not be necessary to depressurize the MR fluid section to perform routine maintenance. ? The part count has been significantly reduced. This redesign was only completed in April, and is shown in Appendix G. Task 10 is, therefore, ongoing.

Phase II Final Report DVMCS p. 14 Task 11: Update economic, market and environmental analyses

Economic & Market Analysis A “bottom up” economic model was prepared for the DVMCS. In this model, typical drilling costs are input for the particular market. Some basic assumptions are made about the improvement provided by the DVMCS, how these savings are to be shared by the service company and its clients, to generate an anticipated revenue per job. The estimated number of jobs per year is also estimated. To estimate costs, we included: cost of purchasing the units; number of units needed to support one job; cost of money; anticipated repair costs and schedule; overhead, etc. On this basis, one can calculate the ROI and payback period for our customer (i.e., the oilfield service or supply company) on purchasing systems. Preliminary calculations show that, with reasonable, conservative assumptions, the DVMCS represents a very attractive investment, even for simple vertical well drilling. In deep, hard rock or offshore drilling the paybacks are enormous. One example of the results of the model is shown in Table 1 , on page 17. The model can be refined as the parameters become better defined during Phase III. To balance this model, we also commissioned a “top down” model for several APS products and potential products. the DVMCS*. This model, produced by Spears & As- sociates, looked at the total domestic drilling market projections, estimated the value of this product, including rig time savings, repair and replacement costs for damaged products, etc. It also assume reasonable sharing of these savings among the end user, service company and supplier (i.e., APS.) It also indicated an enormous market for this product. An excerpt of this study, dealing with the DVMCS, is attached as Appendix H. These studies have been confirmed by the level of customer interest in these tools from several service and supply companies. The results of the testing at TerraTek have only heightened this interest. We anticipate signing one or several non-exclusive agree- ments to provide this DVMCS, or at least the active parts of it, before the end of Phase III. We furthermore anticipate the generation of revenues from the sale or leasing of the “precommercial” prototypes during 2006.

Environmental analysis The DVMCS is, by and large, a standard piece of oilfield hardware. The only non- standard substance is the magnetorheological fluid (MRF). The MRF consists of iron filings suspended in a high-temperature synthetic oil, and therefore poses no more environmental risks than the oil itself, which are minimal. The development of the DVMCS therefore poses no significant environmental risks.

Task 12: Update financing plan The financing plan is now based upon the very likely formation of partnerships with one or more service and supply companies to market the DVMCS. Several have expressed

* Note: This study was part of a general business model and was not charged to this contract.

Phase II Final Report DVMCS p. 15 interest. We fully anticipate that we will earn revenue through the sale or lease of the precommercial prototypes, and this will fully support the efforts of Phase III and beyond. There are several models for the commercialization of the DVMCS, including: ? Construction and sale of the tools by APS Technology. In this model, APS would manufacture the entire tool and sell it to oilfield tool or service companies. ? As an alternative, APS could set up an operation to lease the tools. We are in- vestigating this option for several other of our products. As the tool may be most useful in very specific environments, our customers may prefer this option. ? Sale of the active elements of the DVMCS. Based upon discussions with potential customers, this appears to be the most likely option. The customers would manufacture the conventional components of the DVMCS (bearings, Belleville springs, etc.) Our customers can, in all probability, manufacture these items at lower cost, or adapt their standard parts. APS would provide the active elements, and the software to control them. ? To allow APS to fully share in the potentially enormous market for this tool, a royalty based on usage would be charged, except possibly in the leasing option.

Task 13: Submit Phase II final report This document constitutes the final report on Phase II.

Phase II Final Report DVMCS p. 16 Table 1: Economic Model of DVMCS for Service Company User

REVENUE CALCULATION PER JOB TYPICAL JOB QUANTITY RATE TIME COST ANNUAL

DAYS 25.0 COST PER DAY $ 28,000 TOTAL DRILLING COST $ 700,000 NO OF TRIPS 7.0 HOURS PER TRIP 12.0 TRIPPING HOURS 84.0 MAINTENANCE HOURS/DAY 2.0 MAINTENANCE HOURS 50.0 DRILLING HOURS 466.0

AVD SAVINGS ROP INCREASE 10% DRILLING HOURS SAVED 42.4 TRIPS AVOIDED 2 TRIPPING HOURS SAVED 24.0 NET TIME REDUCTION 66.4 COST SAVINGS $ 77,424 PERCENT TO SERVICE CO 50%

SERVICE CO GROSS REV $ 38,712

NUMBER OF JOBS 12

ANNUAL REVENUE $ 464,545

COST CALCULATION

COST OF TOOL $ 200,000 TOOLS/JOB 2.5 TOTAL TOOL COST $ 500,000 USEFUL LIFE (years) 5 DEPRECIATION (s/l) $ 100,000 INTEREST RATE 6.0% INTEREST COST 30,000 TOOL REFURBISHMENT $ 12,000 DRILLING HOURS/REFURB 500 NUMBER OF REFURBS/YR 11 MAINTENANCE COSTS $ 132,000 DEPLOYMENT COSTS/JOB $ 3,000 TOTAL DEPLOYMENT $ 36,000

TOTAL ANNUAL COST $ 298,000

PAYBACK CALCULATION

NET PROFIT $ 166,545

ROI 33.3% PAYBACK (YEARS) 3.00

Phase II Final Report DVMCS p. 17 Other A paper on the early work on this project was given at the National Gas Technologies II Conference in February, 20042. A paper on this project was also presented to the American Association of Drilling Engineers National Technology Conference in April, 20053. An abstract has been submitted to the IADC/SPE Asia Pacific Drilling Technology Con- ference, 13 - 15 Nov 2006, in Bangkok, Thailand. We also plan to submit an abstract to either the SPE Annual Technical Conference & Exhibition (24-27 Sept 200, San Anto- nio) or to the SPE/IADC Drilling Conference 20-22 February, 2007,Amsterdam.) If this abstract is accepted, we plan to withdraw the one for the APDTC.

Units

To be consistent with standard oilfield practice, English units have been used in this report. The conversion factors into SI units are given below.

1 ft. = 0.30480 m 1 g = 9.82 m/s 1 in. = 0.02540 m 1 klb. = 4448.2 N 1 lb. = 4.4482 N 1 rpm = 0.01667 Hz 1 psi = 6984.76 Pa

References

1 M.E. Cobern, “Downhole Vibration Monitoring & Control System, Phase I Final Report, 31 August, 2004” 2 M.E. Cobern & M.E. Wassell, “Drilling Vibration Monitoring & Control System,” presented at the Natural Gas Technologies II Conference, Phoenix, 8-11 Feb 2004 3 M.E. Cobern & M.E. Wassell, “Laboratory Testing of an Active Drilling Vibration Monitoring & Control System,” presented at the AADE National Technical Conference & Exhibition, Houston, 5-7 April, 2005, paper AADE -05-NTCE -25.

Phase II Final Report DVMCS p. 18 Appendix A: Minimum WOB Variation

Phase II Final Report DVMCS p. 19

advanced product A P S support TECHNOLOGY MEMORANDUM

TO: Marty, Dan, Bill, Doug, Carl

FROM: Mark Wassell

DATE: August 17, 2004

SUBJECT: AVD Sensor Algorithm

CC:

Scope I ran through a number of analyses to get some data for the AVD sensor algorithm. These analyses look at the vibration data that can be easily measure during operation downhole.

Summary 1. It is easy to determine whether the system damping is optimal because the WOB range and the displacement range are minimized. However, when the system does not have optimum damping it is difficult to determine whether there is too much or too little damping. 2. This method uses the absolute linear displacement between the upper and lower housings and the fluctuating WOB to determine whether the damping needs to be adjusted. 3. The WOB measurement needs only to be a relative measurement and therefore does not require the accuracy of the typical WOB tool. Drift, pressure and temperature effects do not need to be included into the measurement, only the range and the average need to be measured. 4. The linear displacement sensor must measure absolute position for this method. 5. One indication that the damping is optimized is that the WOB range is minimal. However, the high the applied weight on bit the greater the WOB range. Therefore without knowing the desired or actual WOB it is difficult to tell whether the system has been optimized. 6. The analysis also shows that when the dynamic spring rate of the system equals the static spring rate the system has been optimized. In general if the dynamic spring rate is greater than the static spring rate, then damping level is to low. If

APS Technology, 800 Corporate Row, Cromwell CT 06416 Phone: 860-613-4450 Fax: 860-613-4455 WEB: aps-tech.com 1 of 5 the dynamic spring rate is less than the static spring rate then the damping is to high. 7. If the minimum displacement is negative then the damping level is to low. However, for high WOB the minimum displacement is positive. Therefore this is only useful at low WOB.

APS Technology, 800 Corporate Row, Cromwell CT 06416 Phone: 860-613-4450 Fax: 860-613-4455 WEB: aps-tech.com 2 of 5

Figure 1 - Damping Control Schematic

APS Technology, 800 Corporate Row, Cromwell CT 06416 Phone: 860-613-4450 Fax: 860-613-4455 WEB: aps-tech.com 3 of 5 6 3/4 AVD WOB Range Variation

600000 Soft Formation - 10,000 lbs WOB Soft Formation - 20,000 lbs WOB Soft Formation - 30,000 lbs WOB 500000 Medium Formation - 10,000 lbs WOB Medium Formation - 20,000 lbs WOB Medium Formation - 30,000 lbs WOB 400000 Hard Formation - 10,000 lbs WOB Hard Formation - 20,000 lbs WOB Hard Formation - 30,000 lbs WOB 300000 WOB Range - lbs - Range WOB 200000

100000

0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Damping lb-sec/in

APS Technology, 800 Corporate Row, Cromwell CT 06416 Phone: 860-613-4450 Fax: 860-613-4455 WEB: aps-tech.com 4 of 5 6 3/4 AVD Stroke Range

40 Soft Formation - 10,000 lbs WOB Soft Formation - 20,000 lbs WOB 35 Soft Formation - 30,000 lbs WOB Medium Formation - 10,000 lbs WOB

30 Medium Formation - 20,000 lbs WOB Medium Formation - 30,000 lbs WOB

25 Hard Formation - 10,000 lbs WOB Hard Formation - 20,000 lbs WOB Hard Formation - 30,000 lbs WOB 20

15 Stroke Range - in

0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Damping lb-sec/in

APS Technology, 800 Corporate Row, Cromwell CT 06416 Phone: 860-613-4450 Fax: 860-613-4455 WEB: aps-tech.com 5 of 5 Appendix B: Hardening Algorithm

Phase II Final Report DVMCS p. 24 AVD – Hardening Damper Scheme

Scope This scheme uses the relative position of the AVD inner housing to the outer housing to set the MR damping. Analysis shows that the greater the travel of the damper the higher the required damping level. Analysis shows that lighter WOB requires less damping than higher WOB. The analysis also shows that the greater the stroke the greater the required damping. The analysis below shows that this concept works for varying amounts of WOB and ROP.

Analysis

The hardening equation:

n c = A x d + B

where: c = damping (lb-sec/in) n A = (dampmax - dampingmin) / disp B = Min damping

AVD Hardening Damper

7000 6000

5000 4000 3000 2000

Damping - lb-sec/in 1000 0 -6 -4 -2 0 2 4 6 Dis p - in

Appendix C: Magnetic Testing Daniel E. Burgess February 25, 2005

Phase II Final Report DVMCS p. 30

OBJECTIVE The magnetic testing was performed to investigate the properties of ferromagnetic materials in a magnet field as relates to the axial valve of the Down Hole Vibration Damper.

TESTING The testing consisted of 3 sets of tests the first was to examine how materials behaved in a magnetic field. The second to test what happens if a strong magnet is placed into a variable shunt devise, and the third testing on the test mandrel used in our DVM valve. The first test involved using a common coil to charge equal length samples with the identical amperages and measuring the resulting magnetic fields. Peak magnetic field strength was recorded with various current settings. Then the coil was turned off and residual field strength was recorded both in and out of the fixture (outside the fixture does not take into account the residual field in the fixture itself). 1018 12L14 4140 Initial reading (kGauss) 0.006 0.108 0.1 Coil at 1 amp (kGauss) 1.7 1.73 1.6 Coil Off (kGauss) 0.35 0.35 0.45 Coil 2 amps (kGauss) 2.5 2.9 2.4 Coil Off (kGauss) 0.39 0.35 0.59 Residual measured outside (kGauss) 0.36 0.14 0.54 Percent of Highest field residual 14% 5% 23% Comments Low residual as measured outside leads me to believe that the short circuit bar was retaining a high enough field to alter off state measurements

Report Title pg. 1

Test Coil The measurement device is pictured here. A steel frame shorts the Gap Spacer poles with a fixed gap at one end maintained by Short Circuit a spacer with a slot that Frame provides space to insert the probe. The coil is a Sample wound boben designed and produced for the Probe RSM solinoids. The current in each test is precisely the same and coil orientation is the same.

This is how residuals Sample were measured outside the frame. A steel desk was used to short the Steel Desk poles with the probe between the desk and the sample, the highest atainable reading was Probe recorded.

The second test performed was to test if it was possible to use the coils to shunt or bias a piece of “magnetically soft” iron to short out the ends of a permanently biased magnetic bar that runs through it’s center. The goal here being that we could use 4140 or some other high strength steel, but most likely having a high residual field, through the center of the valve; and band around it with a magnetically soft material to shunt out that field.

Report Title pg. 2

Above is a sketch of the test setup used. A bar magnet, made of Alnico 5 being 2.5 inches long with an OD of .5 inches, was installed in two separate fixtures. The first fixture made of nylon and having the same OD and ID as the second. The flux was then measured using a loop of HyMU80 a material that has very high permeability and very low coercivity. This approximates the field reaching the OD of the fixture in a very low permeability gap such as if it were in Air. The gauss reading in the “air gap” was 1.365 kGauss. The bar magnet was then placed in the second fixture which comprised a HyMu80 barbell shaped core wrapped with a coil having an OD of 1 inch an ID of .7 inches and a length of 2 inches. The flux was again measured using the same setup as before. (The loop was passed through a demagnetizing coil between each test.) The gauss level through the HyMu80 loop was 3.64 kGauss. The coil was then energized with .5 amps (which should generate approximately 1.1 kGauss) so that the field generated would match the field inside the bar magnet in effect repelling the field out into the loop, which gave it a reading of 5.12 kGauss. It should be noted that the field generated was higher than the sum of individual readings of the magnet and electromagnet. When energized in the opposing direction which should create a short in the bar magnet, the field measured in the HyMu80 loop was 0.00 kGauss. And when power was cut the residual field was 1.34 kGauss which is much less than the original reading, however if the bar magnet is removed and reinstalled the reading is the same as before. The third round of testing consisted of testing the MR coil to try and determine what the residual fields were and how to remove them. The highest residuals were actually found to be in the extension tubes where they neck down close to the mandrel. The shorter of the tubes had a field of 260 Gauss the longer was 350 gauss. Unfortunately these relatively high fields are at the point of greatest impact on the MR fluid (at the entry to the valve). Based on the testing of materials and knowing how the 4140 behaves magnetically we recognized that this area would be very difficult to demagnetize. However we were able to change the geometry of the part to make the strong field have negligible effects on the MR fluid. In the test piece we opened up the bore from .078 clearances to .2575 clearances. This is 3 times more clearance and assuming that the field falls off with the square of the distance this should reduce the field strength to just 10% of it’s original strength. In the new design we have further increased the clearance between the

Report Title pg. 3

mandrel and tubes to .502 inches. So the magnetic field effect should be only about 2.4% of the original field, or 8.5 gauss. The second highest residual fields occurred within the valve itself and were approximately 150 gauss as measured in air (later tests showed this to be 1.4 kGauss in the MR fluid). This is still high enough to thicken the fluid to the point where excessively high damping will be created. This cannot by solved by material selection (as any material will have a residual field) or by geometry because for obvious reasons the valve geometry needs to develop a high field in a tight gap. Thus in order to move to a lower off state residual field you would have to perform some type of demagnetization technique to remove any residual field. With this in mind some tests on the damper section were performed using trace amounts of MR fluid to concentrate the field. In this test we reversed the field applied to the coils but held the current low so that the reversed field exactly matched the residual field. (The 4 amp field used during testing resulted in a residual of approximately 1.4 kGauss; this was equivalent to a 1.26 amp field.) We found that this created a near zero field at the point we had been measuring but other points had residual field of lower magnitude than the initial residual field but with reverse polarity, and still high enough to cause potential problems (approximately 200 gauss) To eliminate these fields we again reversed the polarity of the coil again an fed it an even lower current (.1 amps was found to work well). This canceled out the reverse polarity residual and left us with field that ranged from -20 Gauss to +60 Gauss; this is only 4% of the initial residual field and is not enough to create any appreciable damping. This is how we stumbled upon the stepwise demagnetization technique. We attempted to repeat the results using 120 VAC voltages but found that at 60 Hz the impedance is too great to put the required current through the coil, this meant it was not a viable solution, especially downhole.

ANALYSIS In the first experiment we discovered that different materials will have different field strength when charged by the same coil with the same power. This is the effect of the permeability of the material but also the saturation density. The permeability is similar to resistance; it tells us how easily the field will flow through the material. The saturation density is simply the maximum flux that can be transmitted. In these tests it is very unlikely we got anywhere near the saturation point. A third property is the coercivity, which we had wrongly assumed was a representation of how much residual field would remain after the magnet had been turned off. However in the course of these experiments we found that the residual field even in low coercivity materials can be quite high. Observations during the second test showed that the HyMu80 used in testing was retaining a field, quite a high field of approximately .35 kGauss. Similar to the values attained in testing materials such as 1018 or 4140. However the field was easily reversed. Coercivity is a measure of how easy it is to create and or reverse a field in the material. Fields weaker than the coercivity point will have no magnetic effect on the material. Coercivity is not a measure of maximum residual field, rather it is the minimum field required to develop a response in the material, and therefore represents the lowest possible field that will remain after demaging. The lower the coercivity the material has the easier it is to reverse the field and therefore to demagnetize also weaker fields still affect it so it can be demagnetized to a lower level. Mild steels are not magnetically soft enough for the applications we are attempting hear 1018 has far too high a residual field. The 12L14 (which is a variant of low carbon steel to which lead and phosphorus are added to improve machining) may be a good candidate however finding published magnetic properties on this alloy has proven difficult since it is not considered to be among the “electrical irons” which are optimized for magnetic performance. The Ideal material would have a high permeability, a high saturation density, a low coercive force, and low residual magnetism. Below is a chart published by Carpenter Steels.

Report Title pg. 4

This shows that the best materials for us to use would be the irons and silicon irons. The silicon irons are a bit higher in strength and corrosion resistance and may therefore be the better choice. The experiment with the magnet inside the coil proved interesting. It was possible to create a zero gauss field quite easily, and it was possible to create a very high field without much power. In addition it leads to the belief that we could counter a weak residual field with another weak field of opposite polarity. This conclusion is reached because the residual field after the power reversal is so much lower than the residual field before reverse power was applied. Before applying power the magnetic field was 3.64 kGauss and after reversing the field it dropped to 1.34 kGauss. That means the shunt is conducting approximately 2.3 kGauss preventing it from being exposed to the HyMu loop. This is most likely caused by the internal residual fields within the HyMu shunt continue to be polarized in such a way that it drained off some of the flux preventing it from getting to the test loop. The results from the shunting test lead us to research various demagnetization techniques. A paper by a company called Annis who make demagnetization equipment lists the several methods available. We could heat it past the curie temperature of iron (the Curie temperature is the temperature at which a material looses its ferromagnetism) however the Curie temperature for iron is over 700oC so this option will not work on our valve. Another method is to exactly oppose the field with another field. This would prove difficult since it requires sensors located at strategic points within the mr valve section. It is however very close to what we successfully attempted during the third experiment. The final option is to expose the steel to a field of cyclically reversing polarity with a slowly diminishing intensity. Most commonly an AC field is used, however our findings were that at frequencies that we would get from the alternator the voltage would be insufficient to drive the required current. So we adapted a military technique called “stepwise deperming” that was developed during WWII (and is still used today) to demagnetize submarines. It uses a DC source that is switched in such a way as to reverse the polarity after each cycle and a voltage regulator to reduce the current gradually. We developed a

Report Title pg. 5

circuit to switch DC voltage back and for the while lowering the amplitude continuously. This method uses our available power and does not require sensors in the valve nor much additional circuitry. Our stepwise deperming circuit has been shown to work well during preliminary testing. It was not ready to do a full size test, however we accomplished the same thing manually and were able to bring the sub down to near zero gauss. The system was driven as follows: -4 amps, +3.75, -3.5, +3.25, -3, +2.75, -2.5, +2.25, -2, +1.75, -1.5, +1.25, -1, +0.9, -0.8, +0.7, -0.6, +0.5, -0.4, +0.3, -0.2, +0.1, -0.05, +0.02, -0.01 amps. The entire series of steps took approximately 2 minutes to complete by hand, the circuit running at 10 Hz could do a similar series in 1.25 seconds. A series such as this could be performed at every pumps on / power up to start the tool fresh after every new stand was made up. If a sudden reduction is needed during routine drilling an abbreviated series such as -2, +1, -0.5, +0.1 could be done in a mere fraction of a second.

CONCLUSIONS

1. All materials retain a residual field, but the proportions of residual fields vary from one material to the next and are often not published.

2. Coercivity is not a measure of maximum residual field, rather it is a representation of the minimum field required to develop a response in the material, and therefore represents the lowest field that will remain after demaging.

3. A permanent magnet setup can be used to generate high magnetic fields with low power consumption; however power will be required to go to an off state.

4. A variable shunt device can be used to minimize the effects of residual fields on the fluid but cannot eliminate them without the use of power.

5. Regardless of what material is used, to attain an off state of near zero Gauss we will have to perform a demagnetizing technique

6. The best demagnetizing technique available to us down hole is to periodically reverse the fields applied and gradually lower the field intensity to demagnetize the materials (stepwise deperming).

7. Downhole ready circuits to demagnetize the sub have been successfully built and tested, and we have shown that the technique does indeed work.

Report Title pg. 6 Appendix D: MR Fluid Viscosity Testing Joe Nord February 25, 2005

Phase II Final Report DVMCS p. 37

OBJECTIVE Determination of the optimum mixture of magnetorheological (MR) powder and oil to use in the Active Vibration Damper (AVD) tool. It is important that the MR mixture have a broad range of viscosities that increase as an increasing magnetic field is passed through it. It is desirable for the fluid to be as close to Newtonian as possible meaning that it’s viscosity isn’t affected by shearing. We will also compare the viscosity of the homemade MR fluid with a commercially available Lord brand MR fluid.

TESTING

Equipment Figure 4 shows a model of the test fixture and Figure 2 shows a photo of the apparatus. An electromagnet sends the magnetic field through steel mounting brackets and into the aluminum sample container. A rotational viscometer was used to determine the viscosity. The viscometer spindle was lowered into the sample container, which measures the torque required for the spindle to rotate at different speeds. A vane spindle with 4 blades (Figure 3) was used, which works better than the traditional disc spindle for fluids that have solids suspended in them.

Sample Container

Electromagnet

Figure 1: Diagram of magnetic circuit for sample container

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 1

Figure 2: Photo of the viscometer test apparatus

Figure 3: 4-vane spindle A gaussmeter was used to determine the field through the sample container. Figure 4 below shows the results. At relatively low flux levels the fluid became too viscous to measure. The highest current we could send to the coil was 2 amps, which relates to 240 gauss. Even though the fluid is highly viscous at this flux level, the Lord literature says it is not yet magnetically saturated at this point.

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 2

Flux in Viscometer Test Container

300

250

200

150

100 Magnetic Flux (G)

0 012345 Current (A)

Figure 4: Magnetic flux levels observed in viscometer test container

The rotational viscometer displays three outputs: viscosity, percentage of torque, and rotational speed (RPM). The viscosity is programmed to be accurate when using a 500 mL beaker and a spindle guard. The high cost of the MR fluid and the difficulty involved in putting a magnetic field through a container of this size led to the use of a much smaller sample container (approx 37 mL). Appendix A shows the formulas that were used to convert the torque and rotational speed output of the viscometer into apparent viscosity, shear rate, and shear stress.

Procedures Initial testing on the AVD sub was performed using an 8:1 mass ratio MR fluid. i.e., 8 parts MR powder mixed with 1 part Mobil 624. Viscosity testing was performed with the 8:1 mixture, 6:1, 4:1, 2:1, and a Lord brand silicone-based MR fluid. All of the homemade MR mixtures ran well in the viscometer. The Lord fluid was much more difficult. An important step in determining the apparent viscosity is a plot of the log of the angular velocity versus the log of the torque. For the Lord fluid, these plots were not straight lines. Most sloped downward and then upward while others never even sloped upward. The slope determines the flow behavior index, which is a measure of how it behaves under shear. A flow behavior index above 1 becomes more viscous under shear and an index less than 1 is less viscous under shear. The flow behavior index was so difficult to determine for the Lord fluid that it is unlikely that the calculated apparent viscosity, shear rate, and shear stress are correct.

Plots were created (see Results, below) of the flow behavior index (Figure 5), consistency index (Figure 6), and yield stress (Figure 7) as a function of the magnetic flux in the sample container. The Lord fluid is shown as a dashed line to show that it may not be accurate. The flow behavior index is a measure of how the viscosity reacts to shear. As the MR mixture ratio was increased, the flow behavior index increased and became closer to 1 (closer to Newtonian). All of the flow

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 3

behavior indexes were below 1 so the apparent viscosity reduced as the shear rate increased. The consistency index in Figure 6 is an index of the viscosity at low shear rates. As the MR mixture ratio was increased the consistency index increased. For the magnetic flux levels that we were able to send into the sample cup the consistency index had a broader range as the mixture ratio increased. It is unclear if this would change under higher levels of magnetic flux. The yield stress in Figure 7 is the shear stress required to initiate flow. If the shear stress of the viscometer spindle is lower than the yield stress of the fluid then the spindle will not rotate. The yield stress also has higher values and a broader range as the mixture ratio is increased.

In order to compare the homemade mixtures to the Lord brand MR fluid plots using the raw viscometer data were created. Figure 8 - Figure 11 show the viscosity output versus the rotational speed for different magnetic flux levels. The absolute viscosities are based on using a 500 mL beaker and spindle guard, so they are not accurate. The curves do, however, show good comparative data. For all flux levels, the 6:1 MR mixture ratio plot was closest to the Lord fluid.

RESULTS

0.35

0.3 2 to 1 4 to 1 6 to 1 0.25 8 to 1 LMRF

0.2

0.15 Flow Behavior Index "n"

0.1

0.05

0 0 50 100 150 200 250 300 Magnetic Flux (G)

Figure 5: Comparison of Flow Behavior Index

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 4

450

400

350 2 to 1 4 to 1 6 to 1 300 8 to 1 LMRF

250

200

150 Consistency Index - K (N*s/m^2)

100

0 0 50 100 150 200 250 300 Magnetic Flux (G)

Figure 6: Comparison of Consistency Index

500

450

400

2 to 1 350 4 to 1 6 to 1 300 8 to 1 LMRF

250

Yield Stress (Pa) Stress Yield 200

150

100

0 0 50 100 150 200 250 300 Magnetic Flux (G)

Figure 7: Comparison of Yield Stress

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 5

10000000

1000000

4 to 1 6 to 1 8 to 1 LMRF

100000 Viscometer Viscosity Output

10000 0 20 40 60 80 100 120 Rotational Speed (RPM)

Figure 8: Comparison of Viscometer Output at 0.4A (164 Gauss)

10000000

1000000

4 to 1 6 to 1 8 to 1 LMRF

100000 Viscometer Viscosity Output

10000 0 20 40 60 80 100 120 Rotational Speed (RPM)

Figure 9: Comparison of Viscometer Output at 0.7A (195 Gauss)

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 6

10000000

4 to 1 1000000 6 to 1 8 to 1 LMRF

100000 Viscometer Viscosity Output

10000 0 20 40 60 80 100 120 Rotational Speed (RPM)

Figure 10: Comparison of Viscometer Output at 1 A (218 Gauss)

10000000

4 to 1 6 to 1 8 to 1 LMRF 1000000

100000 Viscometer Viscosity Output

10000 0 2 4 6 8 101214161820 Rotational Speed (RPM)

Figure 11: Comparison of Viscometer Output at 2 A (240 Gauss)

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 7

CONCLUSIONS The fluid with the highest ratio of MR powder had the best viscosity properties. The higher ratio fluids had a broader range of viscosities over the same range of magnetic flux, and they were closer to being Newtonian. However as the ratio is increased the viscosity increases. If a lower viscosity is needed to give the minimum damping in the tool then a lower mixture may be needed. The 6:1 and 4:1 homemade mixtures were the most similar to the Lord brand MR fluid. Lord is highly experienced with MR fluid so they have probably done additional optimization based on the cost of powder, rate of settling out, and other factors. Taking Lord’s experience into account it is recommended that we use the 6:1 ratio fluid in the AVD tool. The 4:1 mixture also had viscosities close to the Lord brand fluid, but 6:1 is more suitable because of its wider range of viscosities.

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 8

APPENDIX A: VISCOSITY CALCULATIONS angular velocity = ω RPM = N

N ω 2.π . 60

A plot of the Log of Torque v. Log of angular velocity determines the flow behavior index (η) and a constant A. The slope is η and the y intercept is the Log of A. This was proven with the calibration fluid. η was equal to 1 which means that it is Newtonian. The constant A provided the correct viscosity 5060 cP using additional equations listed below. For a Newtonian fluid the consistency factor (K) is equal to the viscosity (µ).

flow behavior index = η consistency factor = K

A K η 2 2.π .Rb2.L. 2 η Rb η . 1 Rc Spindle radius = Rb Effective Spindle Length = L Cup radius = Rc

shear rate = SR

2.ω SR 2. 2 η Rb η . 1 Rc

shear stress at the spindle = SS

T SS 2.π .Rb2.L

apparent viscosity = µa . η 1 µa K SR

TR-05-002 Magnetorheological Fluid Viscosity Testing pg. 9

Appendix E: Testing of DVMCS Prototype with Outer Coils

M Wassell Nov 16, 2005

Phase II Final Report DVMCS p. 47 Scope For manufacturing and assembly considerations the MR damper coils have been moved to the outer sleeve. The coil grooves on the inner shaft were filled with a material similar to the shaft, allowing for a continuous magnetic field path. The load range was also increased so that tests could be conducted up to 15,000 lbs WOB. The original tests could only be performed at lower WOB (5,000 lb WOB). Higher loads exceeded the capabilities of the motor and test frame.

Conclusions 1. The results from this test are not as good as for the previous test with the coils mounted on the inner shaft. With the coils mounted on the outer sleeve, the maximum dynamic stiffness is 54,000 lbs / in. The previous tests, with the coils mounted on the inner shaft produced a dynamic stiffness of 120,000 lb/in 2. The dynamic stiffness of the outer coil in the off state is less than the off state for the inner coil. 3. The ratio of stiffness ranges from 2.5 to 4.6 for the new design. The inner coil design ratio is 7 to 10. 4. The dynamic stiffness decreases with the increase in WOB. At 5,000 lbs WOB the dynamic stiffness is 54,000 lbs/in, at 10,000 lbs it drops to 40,000 lbs WOB and at 15,000 lbs WOB it is 26,000 lbs /in. 5. Based on items 1 – 4 above, the gap should be reduced. 6. The dynamic stiffness reduction could be attributed to: Damper gap – The reduced off state dynamic stiffness suggests that the gap could be reduced. This would significantly increase the on state damping, hopefully giving a higher damping ratio. Losses due to filling in the old coil grooves Fixture damping problems with the test fixture. – Each time we set up tests the pneumatic and hydraulic actuators operate differently. The wobble in the graphs could also be attributed to the pneumatic and hydraulic dampers.

Quarterly Progress Report #13 DVMCS p. 11

Discussion Test Parameters • MR Fluid – 6:1 mixture • Gap – 0.031” • Valve materials – 410SS • Cam displacement – 0.708”

Quarterly Progress Report #13 DVMCS p. 12 Outer Coil Results

AVD Dynamic Stiffness 5000 WOB

60000

50000

40000 0.5 Hz 1 Hz 30000 1.5 Hz 2 Hz 20000

Dynamic Stiffness - lb/in 10000

0 012345 Current - Amps

Quarterly Progress Report #13 DVMCS p. 13 AVD Dynamic Stiffness 10,000 WOB

45000 40000 35000 30000 0.5 Hz 25000 1.0 Hz 20000 1.5 Hz 2.0 Hz 15000 10000

Dynamic Damping - lb/in Dynamic Damping 5000 0 012345 Current - Amps

Quarterly Progress Report #13 DVMCS p. 14 AVD Dynamic Stiffness 15000 WOB

30000

25000

20000 0.5 Hz 15000 1.0 Hz 1.5 Hz 10000

Dynamic Stiffness - lb/in Dynamic Stiffness 5000

0 012345 Current - Amp

Quarterly Progress Report #13 DVMCS p. 15 Original Inner Coil Results

AVD Dynamic Stiffness 5000 lbs WOB 140000

120000 /in lb

- 100000 s s e

n 80000 f if t

S 60000 c

i 0.5 Hz m

a 40000 1.0 Hz n

y 1.5 Hz D 20000 2.0 Hz 0 01234 Current - amps

Quarterly Progress Report #13 DVMCS p. 16 Scope For manufacturing and assembly considerations the MR damper coils have been moved to the outer sleeve. The coil grooves on the inner shaft were filled with a material similar to the shaft, allowing for a continuous magnetic field path. The load range was also increased so that tests could be conducted up to 15,000 lbs WOB. The original tests could only be performed at lower WOB (5,000 lb WOB). Higher loads exceeded the capabilities of the motor and test frame.

Conclusions 1. The results from this test are not as good as for the previous test with the coils mounted on the inner shaft. With the coils mounted on the outer sleeve, the maximum dynamic stiffness is 54,000 lbs / in. The previous tests, with the coils mounted on the inner shaft produced a dynamic stiffness of 120,000 lb/in 2. The dynamic stiffness of the outer coil in the off state is less than the off state for the inner coil. 3. The ratio of stiffness ranges from 2.5 to 4.6 for the new design. The inner coil design ratio is 7 to 10. 4. The dynamic stiffness decreases with the increase in WOB. At 5,000 lbs WOB the dynamic stiffness is 54,000 lbs/in, at 10,000 lbs it drops to 40,000 lbs WOB and at 15,000 lbs WOB it is 26,000 lbs /in. 5. Based on items 1 – 4 above, the gap should be reduced. 6. The dynamic stiffness reduction could be attributed to: ? Damper gap – The reduced off state dynamic stiffness suggests that the gap could be reduced. This would significantly increase the on state damping, hopefully giving a higher damping ratio. ? Losses due to filling in the old coil grooves ? Fixture damping problems with the test fixture. – Each time we set up tests the pneumatic and hydraulic actuators operate differently. The wobble in the graphs could also be attributed to the pneumatic and hydraulic dampers.

Phase II Final Report DVMCS p. 48 Discussion Test Parameters ? MR Fluid – 6:1 mixture ? Gap – 0.031” ? Valve materials – 410SS ? Cam displacement – 0.708”

Phase II Final Report DVMCS p. 49 Outer Coil Results

AVD Dynamic Stiffness 5000 WOB

60000

50000

40000 0.5 Hz 1 Hz 30000 1.5 Hz 2 Hz 20000

Dynamic Stiffness - lb/in 10000

0 0 1 2 3 4 5 Current - Amps

Phase II Final Report DVMCS p. 50 AVD Dynamic Stiffness 10,000 WOB

45000 40000 35000 30000 0.5 Hz 25000 1.0 Hz 20000 1.5 Hz 2.0 Hz 15000 10000

Dynamic Damping - lb/in 5000 0 0 1 2 3 4 5 Current - Amps

Phase II Final Report DVMCS p. 51 AVD Dynamic Stiffness 15000 WOB

30000

25000

20000 0.5 Hz 15000 1.0 Hz 1.5 Hz 10000

Dynamic Stiffness - lb/in 5000

0 0 1 2 3 4 5 Current - Amp

Phase II Final Report DVMCS p. 52 Original Inner Coil Results

AVD Dynamic Stiffness 5000 lbs WOB 140000

120000

100000

80000

60000 0.5 Hz 40000 1.0 Hz 1.5 Hz Dynamic Stiffness - lb/in 20000 2.0 Hz 0 0 1 2 3 4 Current - amps

Phase II Final Report DVMCS p. 53 Appendix F: Analysis of TerraTek Results

Mark E. Wassell April 16, 2006

Phase I Final Report DVMCS pg. 54 Scope Drilling tests of the Downhole Vibration Monitor & Control System (DVMCS), also re- ferred to as the Active Vibration Damper (AVD), were performed at the TerraTek test facility in Salt Lake City. The drilling tests consisted of drilling a number of holes through concrete and granite test blocks. The tests blocks were constructed such that they would develop drilling vibrations. A one-foot thick granite block was sandwiched between an upper and lower section of concrete. The granite was mounted within the blocks at a 10o angle. This arrangement was intended to develop increased vibration as the rock bit drill through the interface. A total of four test blocks were constructed. Each test block could accommodate 8 holes. The drilling simulator at TerraTek was used to drill the holes. The test rig could accommodate only a short drilling assembly, so the effects of the mass and length of the drillstring could not be properly accounted for in the tests. This also meant that the drilling vibration environment was fairly benign.

The intent of the tests were to: ? Determine the drilling performance of the AVD tool under various loading conditions. ? Compare the performance of the AVD to that of a standard drill collar ? Evaluate the bearings and seals under drilling conditions

A number of tests were performed while drilling each hole. Variables included: ? Formation material: ? Rotary speed ? Weight on bit (WOB) ? AVD damping

Conclusions 1. The drilling at TerraTek was fairly benign. The vibration levels were low and the changes in WOB were minor. Testing did not include any bit bounce. The Ter- raTek test rig, with its short BHA section, is ideal for drilling with standard collars. The short length minimizes axial vibration due its stiffness and high fundamental natural frequency. 2. Significant bit wear occurred during the tests. For a standard collar the ROP at the end of the tests was 60% of the ROP at the start of the tests. To account for this the ROP rates were corrected based on a linear degradation of the bit for

Phase I Final Report DVMCS pg. 55 each hole drilled. Other drilling results, such as WOB, TOB acceleration, have not been corrected for bit wear. 3. Even with the benign drilling conditions the AVD showed its ability to improve ROP (Figure 11). While drilling through concrete the AVD improved ROP by 10 – 15%. For granite the improvement was up to 11%. While drilling through granite at 120rpm with 15,000 WOB the ROP actually dropped by 2%. However, lighter damping may improve this situation. The lighter damping case was not run. 4. The optimum drilling condition occurs when the AVD internal travel is ~0.17”. If the drilling becomes too smooth, the ROP actually drops. There appears to be an optimum travel and WOB fluctuation that produces the best ROP. 5. The vibration levels measured through the tests were fairly benign. These were in the 5 – 25 g level. 6. The AVD tool significantly decreased the WOB fluctuations compared to the standard drill collar (Figure 13). The AVD reduced the WOB fluctuation by 60%. This should be even more beneficial under high vibration conditions. 7. The AVD displacement was 0.25” for the power off state and 0.13” at full power. The dynamic stiffness varied from 4200 lb/in to 17,800 lb/in. These ranges should significantly improve for the commercial tool. Upon disassembly of the AVD tool, it was found that the brass bobbins for the damper coils had become slightly magnetic. This would have reduce thed performance of the magnetic coils. The tool was disassembled after the completion of the tests to inspect the seals and bearings. Both the bearings and the seals were still in good working order.

Recommendations 1. The AVD tool should be run in a test well under drilling conditions that would develop significant WOB fluctuations, including bit bounce. 2. An algorithm based on optimum tool displacement for a given acceleration should be developed and programmed into the tool. 3. The tool should also be programmed to perform stepped damping level sweeps during the testing. The tool should record displacement and acceleration and be time tagged for reference to ROP, rotary speed, WOB and formation.

Discussion The AVD tests were performed at the TerraTek facility in Salt Lake City. TerraTek has a drilling simulator that is primarily used for testing bits. (See Figure 9, below.) The formation block is mounted in a pit beneath the drilling rig. The maximum BHA length that the drilling simulator can accommodate is approximately 30 feet. The WOB load is

Phase I Final Report DVMCS pg. 56 applied using hydraulic rams. This makes for a very stiff drilling structure, much stiffer than conditions that will occur under normal drilling, especially in the deep, hard rock drilling conditions for which the AVD was designed.

Figure 9: Layout of TerraTek Drilling Facility (courtesy of TerraTek, Inc.)

Special concrete and granite layered structures were built for the AVD tests. These were intended to induce as much axial drilling vibration as possible. To accomplish this the granite slab was placed on a 10o angle between a top and bottom layer of concrete. This created an interrupted cut at the interface between the layers. Four tests blocks were built in which 8 holes each were drilled.

Baseline reference points were developed for the various conditions using a conventional drill collar in lieu of the AVD. At the completion of the AVD tests, the baseline was rerun to determine the affect of bit wear on ROP (Figure 10).

Phase I Final Report DVMCS pg. 57 Figure 10

ROP CORRECTION FACTOR

1.2

0.8

0.6

0.4 Concrete

ROP FACTOR 0.2 Granite

0 0 10 20 30 HOLE NUMBER

The ROP results for the AVD tool were corrected using Figure 10 based on the drilled hole number. Figure 10 shows that there was significant bit wear during the week long drilling tests.

A number of tests were performed in each of the holes varying the formation material, rotary speed, WOB and damping. The layered formation blocks gave four different formation scenarios for drilling: ? Concrete ? Concrete into granite ? Granite ? Granite into concrete ? A second layer of concrete

Holes were drilled through the different sections with different rotary speeds, WOB and AVD damping levels. The data recorded for each drilled section included: ? ROP ? TOB ? Acceleration ? AVD displacement The data was sampled at frequency of 2000 samples per second.

Phase I Final Report DVMCS pg. 58 Tests were also performed using an algorithm designed to optimize the AVD performance. The algorithm was developed based on analytical analysis for hard rock drilling. The algorithm was modified during the tests to improve the performance. The electron- ics feedback system broke down during the algorithm tests. Because the drilling conditions were benign and did not change significantly during testing, the algorithm could be approximated by adjusting the damping levels manually.

Phase I Final Report DVMCS pg. 59 Results

Key: In each of these figures, the curves are labeled in the following manner: M-rrr-ww, where: ? M is the material being drilled – Concrete or Granite ? rrr is the rotary speed in rpm ? ww is the weight-on-bit (WOB) in klbs.

Furthermore, concrete data are shown as open points and common RPM-WOB combi- nations are shown by the same shape data points.

Phase I Final Report DVMCS pg. 60 Figure 11

DVMCS TerraTek Test ROP Relative Improvement*

1.25

C-120-10 C-120-15 G-120-10 G-120-15 G-180-10 G-180-15 1.00 ROP Improvement Factor

*Relative to baseline conventional collar 0.75 0 33 67 100 Damping - Percent Full Power

Phase I Final Report DVMCS pg. 61 Figure 12

DVMCS TerraTek Test Uphole Relative Acceleration*

1.5 C-120-5 C-120 -10 G-120-10 G-120-15 G180 -10 G-180-15

1.0 Relative Acceleration

* Relative to baseline conventional collar 0.5 0 33 67 100 Damping - Percent of Full Power

Phase I Final Report DVMCS pg. 62 Figure 13

DVMCS TerraTek Test Relative WOB Variation*

1.0

0.8

0.6

0.4 C-120-5 C-120-10 G-12O-5

Relative WOB Range *Relative to baseline conventional collar G-120-15 0.2 G-180-10 G-180-15

0.0 0 33 67 100 Damping - Percent Full Power

Phase I Final Report DVMCS pg. 63 Figure 14

DVMCS TerraTek Test Uphole Relative Acceleration*

1.0

C-120 -10 C-120 15 G-120-10 G-120-15 G-180-10 Relative Acceleration G-180-15

*Relative to baseline conventional collar

0.5 0 33 67 100 Damping - Percent of Full Power

Phase I Final Report DVMCS pg. 64 Figure 15

DVMCS TerraTek Test Mandrel Displacement

0.3

0.2

C-120-10 C-120-15 0.1 G-120-10 Mandrel motion (in.) G120-15 G-180-10 G-180-15

0.0 0 33 67 100 Damping - Percent of Full Power

Phase I Final Report DVMCS pg. 65 Figure 16

DVMCS TerraTek Test Dynamic Stiffness

20.0 C-120-10 C-120 -15 G-120 -10 G-120-15 15.0 G-180-10 G-180-15 lbs./in (Thousands) 10.0

5.0 Dynamic Stiffness

0.0 0 33 67 100 Damping - Percent of Full Power

Phase I Final Report DVMCS pg. 66

WOB TOB ACC HOLE HOLE # MATERIAL RPM WOB COLLAR DAMPING FREQ MAX MIN AVE RANGE FREQ MAX MIN AVE RANGE ROP MAX MIN FREQ 2-1 CONCRETE 120 5000 COLLAR NA 2 5388 4094 4741 1294 40 835 -506 164.5 1341 10 8 -7 2-2 CONC-GRAN 120 5000 COLLAR NA 3 6288 4100 5194 2188 44 941 -618 161.5 1559 8 11 -9 2-3 GRANITE 120 5000 COLLAR NA 2 5388 4329 4858.5 1059 40 786 -689 48.5 1475 4 12 -11 2 2-4 GRANITE 220 5000 COLLAR NA 5 5271 4541 4906 730 44 976 -506 235 1482 5 20 -12 2-5 GRANITE 220 10000 COLLAR NA 5 10600 9494 10047 1106 - 1323 -436 443.5 1759 12 32 17 2-6 GRANITE-CONCRETE 120 5000 COLLAR NA 2 5553 4565 5059 988 45 865 -560 152.5 1425 3 12 12 3-1 CONCRETE 180 5000 COLLAR NA 3 5576 4306 4941 1270 45 1118 -506 306 1624 17 13 -6 3-2 CONCRETE-GRANITE 180 5000 COLLAR NA 3 6218 4241 5229.5 1977 42 1043 -571 236 1614 21 11 -9 3-3 3 GRANITE 180 5000 COLLAR NA 3 5435 4447 4941 988 42 982 -520 231 1502 5 21 -10 3-4 GRANITE 180 10000 COLLAR NA 3 10941 9353 10147 1588 42 1348 -553 397.5 1901 10 23 -13 3-5 GRANITE-CONCRETE 180 5000 COLLAR NA 3 5459 4565 5012 894 42 992 -469 261.5 1461 4 19 -13 4-1 CONRETE 120 10000 COLLAR NA 6 11106 9176 10141 1930 46 1294 -471 411.5 1765 37 10 -10 4-2 CONCRETE-GRANITE 120 10000 COLLAR NA 1 11612 9106 10359 2506 28 1224 -682 271 1906 8 15 -11 4-3 4 GRANITE 120 10000 COLLAR NA 1 10788 9071 9929.5 1717 45 1082 -682 200 1764 7 16 -15 4-4 GRANITE 120 15000 COLLAR NA 1 15859 13918 14888.5 1941 42 1424 -471 476.5 1895 14 18 -13 4-5 GRANITE-CONCRETE 120 10000 COLLAR NA 2 10635 9129 9882 1506 44 1012 -471 270.5 1483 7 15 -11 5-1 CONCRETE 180 10000 COLLAR NA 3 11200 8706 9953 2494 42 1218 -759 229.5 1977 54 14 -8

Collar 5-2 CONCRETE-GRANITE 180 10000 COLLAR NA 3 10694 9282 9988 1412 42 1259 -682 288.5 1941 24 23 -11 5 5-3 GRANITE 180 10000 COLLAR NA 1 10737 9250 9993.5 1487 43 1424 -718 353 2142 10 23 -13 5-4 GRANITE-CONCRETE 180 10000 COLLAR NA 1 10765 9282 10023.5 1483 43 1412 -759 326.5 2171 10 24 -16 6-1 CONCRETE 120 15000 COLLAR NA 1 14247 11612 12929.5 2635 42 1471 -259 606 1730 70 11 -10 6-2 CONCRETE-GRANITE 120 15000 COLLAR NA 2 16176 13647 14911.5 2529 2 1559 -700 429.5 2259 13 15 -15 6 6-3 GRANITE 120 15000 COLLAR NA 2 15824 14024 14924 1800 42 1398 -598 400 1996 12 15 -15 6-4 GRANITE-CONCRETE 120 15000 COLLAR NA 2 16388 14129 15258.5 2259 43 1424 -429 497.5 1853 12 12 -14 7-1 CONCRETE 180 15000 COLLAR NA 0 0 7-2 CONCRETE-GRANITE 180 15000 COLLAR NA 3 16294 14353 15323.5 1941 3 1794 -606 594 2400 21 25 -18 7 7-3 GRANITE 180 15000 COLLAR NA 3 15812 14424 15118 1388 3 1553 -547 503 2100 19 21 -15 7-4 GRANITE-CONCRETE 180 15000 COLLAR NA 1 15941 11353 13647 4588 42 1553 -382 585.5 1935 22 20 -13 8-1 CONCRETE 90 20000 COLLAR NA 4 21706 18529 20117.5 3177 31 1965 -12 976.5 1977 69 6 -10 8-2 CONCRETE-GRANITE 90 20000 COLLAR NA 2 23753 17576 20664.5 6177 31 2345 -185 1080 2530 62 11 -11 8 8-3 GRANITE 90 20000 COLLAR NA 2 21247 18753 20000 2494 48 1718 -300 709 2018 17 13 -13 8-4 GRANITE-CONCRETE 90 20000 COLLAR NA 2 22500 17453 19976.5 5047 48 2324 -324 1000 2648 30 13 -14

Phase II Final Report DVMCS p. 67 WOB TOB ACC HOLE HOLE # MATERIAL RPM WOB COLLAR DAMPING FREQ MAX MIN AVE RANGE FREQ MAX MIN AVE RANGE ROP MAX MIN FREQ 4-1 CONCRETE 120 15000 AVD 0 4 15200 14141 14670.5 1059 20 1794 -182 806 1976 48.9 12 -8 6 4-1 CONCRETE 120 15000 AVD 1.257 4 15200 14141 14670.5 1059 20 1794 -182 806 1976 58.8 12 -8 6 4-2 12 CONCRETE-GRANITE 120 15000 AVD 0 1 15271 14376 14823.5 895 43 1512 -629 441.5 2141 15.41 17 -14 136 4-3 GRANITE 120 15000 AVD SWEEP 2 15506 14329 14917.5 1177 43 1677 -516 580.5 2193 11.2 18 -13 4 4-4 GRANITE-CONCRETE 120 15000 AVD 0 2 15412 14471 14941.5 941 48 1512 -465 523.5 1977 10.24 17 -12 6 5-1 CONCRETE-GRANITE 180 15000 AVD 0 3 15506 14471 14988.5 1035 42 1559 -371 594 1930 20.65 20 -11 5-2 13 GRANITE 180 15000 AVD SWEEP 3 15506 14235 14870.5 1271 42 1746 -549 598.5 2295 14.71 22 -15 211 5-3 GRANITE-CONCRETE 180 15000 AVD SWEEP 3 15647 14518 15082.5 1129 42 1512 -371 570.5 1883 16.32 21 -11 210 6-1 CONCRETE 120 5000 AVD SWEEP 2 5176 4424 4800 752 2 1029 -647 191 1676 7.65 14 -11 11 6-2 CONCRETE-GRANITE 120 5000 AVD 1.335 3 5435 4541 4988 894 44 900 -647 126.5 1547 12.24 12 -10 11 6-3 14 GRANITE 120 5000 AVD 1.335 16 5082 4353 4717.5 729 35 829 -512 158.5 1341 3.3 12 -8 28 6-4 GRANITE 120 10000 AVD SWEEP 3 10224 9447 9835.5 777 42 1300 -594 353 1894 4.05 14 -11 7 6-5 GRANITE-CONCRETE 120 10000 AVD 1.335 2 10271 9376 9823.5 895 42 1176 -559 308.5 1735 5 14 -8 11 7-1 CONCRETE 120 10000 AVD 2.66 2 10865 9065 9965 1800 36 1300 -388 456 1688 33.53 13 -11 105 7-2 CONCRETE-GRANITE 120 10000 AVD 2.66 3 11182 8959 10070.5 2223 37 1465 -562 451.5 2027 34.7 14 -10 11 15 7-3 GRANITE 120 10000 AVD 2.66 3 10129 9259 9694 870 41 1176 -718 229 1894 5.3 14 -11 7 7-4 GRANITE-CONCRETE 120 10000 AVD 2.66 2 10341 9235 9788 1106 45 1238 -547 345.5 1785 5.76 13 -12 11 8-1 CONCRETE 120 10000 AVD 4 2 10588 8912 9750 1676 45 1382 -471 455.5 1853 32.8 11 -7 11 8-2 CONCRETE-GRANITE 120 10000 AVD 4 2 11500 9135 10317.5 2365 36 1176 -388 394 1564 25.74 11 -9 211 8-3 16 GRANITE 120 10000 AVD 4 2 10341 9353 9847 988 32 1021 -447 287 1468 4.55 12 -9 11 8-4 GRANITE-CONCRETE 120 10000 AVD 4 8-4.5 GRANITE 120 15000 AVD 4 15412 14235 14823.5 1177 1341 -347 497 1688 9.34 12 -11 9-1 CONCRETE 120 15000 AVD 1.33 1 16294 13706 15000 2588 46 1676 -259 708.5 1935 59.47 12 -8 6 9-2 CONCRETE-GRANITE 120 15000 AVD 1.33 2 15694 14424 15059 1270 40 1553 -382 585.5 1935 12.53 15 -9 6 17 9-3 GRANITE 120 15000 AVD 1.33 2 15412 14471 14941.5 941 43 1418 -465 476.5 1883 10.01 12 -12 11 9-4 GRANITE-CONCRETE 120 15000 AVD 1.33 Not good data 10-1 CONCRETE 120 15000 AVD 2.66 16341 13706 15023.5 2635 1653 -229 712 1882 63.38 12 -9 10-2 CONCRETE-GRANITE 120 15000 AVD 2.66 1 16435 13941 15188 2494 38 1606 -606 500 2212 68.72 12 -11 5 18 10-3 GRANITE 120 15000 AVD 2.66 15482 14424 14953 1058 1276 -465 405.5 1741 9.11 12 -12 10-4 GRANITE-CONCRETE 120 15000 AVD 2.66 15294 14518 14906 776 1229 -559 335 1788 8.51 12 -11 11-1 CONCRETE 120 15000 AVD 4 16153 13894 15023.5 2259 1718 29 873.5 1689 55.76 9 -11 AVD 11-2 CONCRETE-GRANITE 120 15000 AVD 4 16153 13579 14866 2574 1718 29 873.5 1689 74.34 9 -11 19 11-3 GRANITE 120 15000 AVD 4 15529 14424 14976.5 1105 1429 -259 585 1688 9.51 13 -14 11-4 GRANITE-CONCRETE 120 15000 AVD 4 2 15412 14541 14976.5 871 35 1435 -224 605.5 1659 8.5 12 -11 11 12-1 CONCRETE 180 15000 AVD 1.33 16106 13941 15023.5 2165 1965 -94 935.5 2059 90.99 16 -12 12-2 CONCRETE-GRANITE 180 15000 AVD 1.33 15412 14235 14823.5 1177 1512 -321 595.5 1833 15.44 21 -13 20 12-3 GRANITE 180 15000 AVD 1.33 15318 14471 14894.5 847 1435 -224 605.5 1659 14.03 18 -14 12-4 GRANITE-CONCRETE 180 15000 AVD 1.33 15247 14329 14788 918 1529 -206 661.5 1735 14.59 15 -14 15-1 CONCRETE 180 15000 AVD ALGOR-1 Bad data - Excessive WOB - No ROP data 15-2 23 CONCRETE-GRANITE 180 15000 AVD ALGOR-1 Bad data - No data in the file 15-3 GRANITE 180 15000 AVD ALGOR-1 Bad data - No data in the file 16-1 CONCRETE 180 10000 AVD ALGOR-2 10271 9212 9741.5 1059 1365 -294 535.5 1659 52.99 14 -5 16-2 CONCRETE-GRANITE 180 10000 AVD ALGOR-2 10906 9400 10153 1506 1329 -329 500 1658 51.94 15 -4 24 16-3 GRANITE 180 10000 AVD ALGOR-2 10224 9494 9859 730 1029 -294 367.5 1323 5.76 15 -8 16-4 GRANITE-CONCRETE 180 10000 AVD ALGOR-2 10200 9471 9835.5 729 1047 -365 341 1412 5.24 18 -8 17-1 CONCRETE 180 5000 AVD 0.158 5294 4400 4847 894 971 -335 318 1306 19.81 12 -4 17-2 CONCRETE-GRANITE 180 5000 AVD 0.158 5506 4282 4894 1224 971 -441 265 1412 22.6 14 -9 17-3 GRANITE 180 5000 AVD 0.158 5294 4376 4835 918 882 -471 205.5 1353 3.56 12 -8 17-4 24.5 GRANITE 240 5000 AVD 0.158 5412 4306 4859 1106 926 -241 342.5 1167 3.19 21 -7 17-5 GRANITE 240 10000 AVD 0.158 10271 9282 9776.5 989 1100 -194 453 1294 9.04 24 -9 17-6 GRANITE-CONCRETE 240 5000 AVD 0.158 5365 4259 4812 1106 912 -312 300 1224 3.88 19 -11 17-7 GRANITE-CONCRETE 180 5000 AVD 0.158 Bad Data 18-1 CONCRETE-GRANITE 180 15000 AVD 2.66 15459 14353 14906 1106 1294 -612 341 1906 16.22 20 -8 18-2 25 GRANITE 180 15000 AVD 2.66 15506 14424 14965 1082 1382 -471 455.5 1853 12.38 20 -10 18-3 GRANITE-CONCRETE 180 15000 AVD 2.66 16012 13518 14765 2494 1465 -265 600 1730 53.08 18 -8 19-1 CONCRETE-GRANITE 180 15000 AVD 4 15529 14424 14976.5 1105 1471 -424 523.5 1895 12.58 21 -8 19-2 26 GRANITE 180 15000 AVD 4 15365 14424 14894.5 941 1506 -224 641 1730 11.62 19 -12 19-3 GRANITE-CONCRETE 180 15000 AVD 4 9 18425 11386 14905.5 7039 43 2041 -218 911.5 2259 22.05 19 -7 10 20-1 CONCRETE-GRANITE 180 10000 AVD 1.33 10412 9447 9929.5 965 1135 -429 353 1564 8.51 18 -4 20-1A CONCRETE-GRANITE 180 10000 AVD 0-1.33 10506 9824 10165 682 1053 -224 414.5 1277 14.93 20 -6 27 20-2 GRANITE 180 10000 AVD 1.33 10224 9565 9894.5 659 1135 -229 453 1364 5.09 16 -7 20-3 GRANITE-CONCRETE 180 10000 AVD 1.33 10224 9376 9800 848 1153 -329 412 1482 4.54 18 -4 21-1 CONCRETE-GRANITE 180 10000 AVD ??? 10329 8953 9641 1376 1341 -265 538 1606 34 17 -4 21-2 28 GRANITE 180 10000 AVD 2.66 10341 9471 9906 870 1128 -391 368.5 1519 6.45 18 -9 21-3 GRANITE 180 10000 AVD 4 10435 9518 9976.5 917 1206 -324 441 1530 5.6 20 -9

Phase II Final Report DVMCS p. 68 Appendix G: Final Design of Field Prototype

Phase II Final Report DVMCS p. 69 REVISIONS REV DESCRIPTION DRAWN CHECKER TURBINE ALTERNATOR TFR CAP A MARKETING RELEASE 4.5 IF, (NC50) SUB 04/27/06 04/27/06 BOX TURBINE ALTERNATOR FEED THRU

UPHOLE CONTROL ASSEMBLY

LVDT AND COMP ACTIVE MAGNET SUB SUB

MR VALVE ASSEMBLY

SPRING STACK TORSIONAL BEARING SUB SUB 4.5 IF, (NC50) PIN

DOWN HOLE

SPRING & BEARING ASSEMBLY COMPANY CONFIDENTIAL PROPRIETARY & TRADE SECRET INFORMATION APPROVALS APS TECHNOLOGY ANY USE OR REPRODUCTION OF THIS DOCUMENT, BY DATE 800 CORPORATE ROW IN PART OR WHOLE, WITHOUT EXPRESS WRITTEN DRAWN RZASA 04/27/06 PERMISSION OF APS TECHNOLOGY IS PROHIBITED CROMWELL, CT 06416−2072 CHECKED CAP 04/27/06 UNLESS OTHERWISE SPECIFIED ENGRG DEB 04/27/06 ACTIVE VIBRATION DAMPER MFG BWP MATERIAL: HEAT TREAT: FINISH: DIMENSIONS ARE IN INCHES & APPLY AFTER FINISH 04/27/06 ASSEMBLY PART MUST BE FREE OF BURRS AND/OR FLASH TOLERANCES N/A N/A N/A BREAK SHARP EDGES APPROX .005 .XXX ±.005 FRACTIONS ±1/64 DWG NO REV FILLET RADII .020 MAX .XX ±.01 ANGLES ±2° B SK−10561 A 63 INTERPRET DWG PER ASME Y14.5M−1994 DIMS IN PARENTHESIS ARE REF ONLY APS FILE : SURFACE FINISH MAX SCALE: 1/10 SHEET 1 OF 1 SK−10561 A ACTIVE VIBRATION DAMPER ASSEMBLY(MARKETING).SLDDRW DO NOT SCALE DWG

PROJECT # XXX−XXX Appendix H: Excerpt from Spears Market Study

Phase II Final Report DVMCS p. 71

Markets and demand for new products from APS Technology

12 July 2004

Prepared by

Richard Spears Spears & Associates, Inc. 5110 South Yale, Suite 410 Tulsa, Oklahoma 74135 918/496-3434 [email protected]

Introduction

APS Technology has several technologies in various stages of development – some in the earliest phases of design, others in the initial steps of production. Spears & Associates, along with Angie Smith and Derek Barnes (APS), prepared this initial evaluation of the addressable market1 for each of seven technologies.

The technologies fall in two categories:

Drilling Technologies

o High temperature turbine alternator o Low cost MWD o Active vibration dampener o Rotary steerable motor o Survey-While-Drilling o Automatic Driller

Production Technologies

o Petromax

Six of the technologies are systems which, when taken to the wellsite by the appropriate company, can be sold as a service to the end user – the oil company. One – the high temperature turbine alternator – is a component for high-end electrical systems used for downhole logging, drilling or other services.

Summary

Based on an analysis of the markets these seven technologies are most likely fit and on a projection of those markets based on activity drivers, the table below outlines our estimate of the total addressable market for each product:

APS-addressable markets by product

Product Addressable Market Units 2003 2004 2005 2006 2007

HT Turbine Alternator High End LWD & RS Systems 1080 1200 1240 1350 1430 Low Cost MWD Footage (Millions) 55 61 60 62 64 Active Vibration Dampener Footage (Millions) 186 206 200 204 207 Rotary Steerable Motor Footage (Millions) 90 100 97 100 102 Survey While Drilling Footage (Millions) 152 168 163 164 166 Automatic Driller Footage (Millions) 3 3 3 3 3 Petromax Flow measurement spending (Millions) $9 $11 $13 $15 $18

APS will prepare estimates of market penetration and resulting sales revenue in a separate document. On the following pages are discussions of the drivers of demand and short evaluations of each technology.

1 By “addressable market” we mean that portion of the market – drilled footage, logging systems, etc. – that an APS product could reasonably expect to serve given the right technology, the right price and the right service company.

Drivers of demand

Drilling Technologies

Demand for drilling technologies is generally driven by the following factors:

o Drilling activity o Directional drilling activity o Revenues related to directional drilling and LWD services

Global drilling activity has remained fairly flat since 1990, with little change through 2007, but directional drilling footage during that same period has been growing strongly, although Spears’ forecast predicts a downward correction in 2005:

Drilling Rigs Directional Footage 2500 110

100 Offshore 2000 90 Land 80

1500 70 60

Rigs 50 1000 Millions 40 30 500 20

0 0

Some of the products APS is considering developing also serve the vertical drilling market. Survey-While-Drilling, for example, is targeted at the vertical section of any hole, whether on land or offshore – even the directionally drilled offshore wells usually have vertical hole sections. As the chart below shows, vertical footage ties more closely to rig count than directional footage:

Vertical Footage 250

Of f shore 200 Land

150

100

3 APS Technology has an automatic or self-propelled driller that appears to fit well with coiled tubing drilling applications. BP has kept two or more CTD units employed for the last several years on thru-tubing re-entry work on the North Slope. As the charts below indicate, about 800 wells per year are drilled with coiled tubing in North America (the largest CTD market by far) and about 85% of the wells are in Canada. These Canadian wells are, for the most part, shallow, straight holes drilled from the surface to TD:

Location of CTD Wells in North Annual Coiled Tubing Drilled Wells Worldwide 1000 America Other 11% 800

Alaska 600 4%

Wells 400

200

0 Canada 85%

1990 1992 1994 1996 1998 2000 2002 2004 2006

Another indicator of potential opportunity for the automated driller is the number of re- entry wells being drilled in mature fields today. No information is available for the rest of the world, but the chart below of re-entries in the US is a good indicator of the trend throughout the industry:

Re-Entry Wells - US 3000

2500

2000

1500 Wells 1000

500

0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Re-Entry 778 971 1469 1097 918 1744 1939 1831 2176 2486 2220 2397

APS Technology’s new products can address several types of drilling applications, from extreme condition wells to shallow, low cost wells, but none address exactly the same sub-divisions of the global drilling market. Therefore, we have divided the global drilling market into four segments, or sub-divisions, to more properly describe the wells and drilling applications each product will serve. The graphic below for 2003 shows the approximate drilling and completion spending found in each sub-division and the approximate number of feet of hole drilled – the size of each box is proportioned to drilling and completion spending in that segment:

Global Drilling Market Segmented by Cost & Difficulty of Drilling Footage – 305 million feet D&C Spending - $80 billion High

104 M ft 34 M ft $16 B $32 B

II IV

Technical Difficulty I III

74 M ft 93 M ft $11 B $21 B

Low High Drilling & Completion Cost

Drilling in the four segments can be described as follows:

SEGMENT I: Shallow land drilling as is typically found in North America, South America and other <5000’ applications. Also includes some shallow water development drilling. These holes have low rig costs and are technically simple to drill.

SEGMENT II: More technically difficult wells, but still with fairly low drilling costs, such are found in the horizontal drilling plays of the Austin Chalk or in certain Middle East and African provinces.

SEGMENT III: Can include offshore development drilling from jackups, like the Gulf of Mexico Shelf, deep land drilling and some international offshore work. Rig costs are higher, but well profiles are still technically simple.

SEGMENT IV: These high cost, technically challenging wells include all deepwater drilling, high temperature/high pressure drilling and deep GOM Shelf work. Also includes remote or international exploration operations.

During 2003, Segment I drilling represented 74 million feet of hole drilled and about $11 billion in drilling and completion spending while Segment IV drilling was 34 million feet and $32 billion in D&C spending. All four sub-divisions together saw 305 million feet of hole drilled and completed at a cost of $80 billion.

Looking forward, with international drilling rising and North American drilling peaking, the overall forecast of drilling through 2007 is fairly flat, although 2004 will be up 10% over 2003. The table below records Spears’ forecast of drilling activity (footage) by segment and by directional vs. vertical, for the period 2003-20072:

Footage Drilled by Segment (Millions) 2003 2004 2005 2006 2007

SEG I Directional 16 18 17 18 19 SEG II Directional 34 38 37 38 39 SEG III Directional 28 31 30 31 32 SEG IV Directional 10 11 11 12 13

Sub Total 88 97 95 99 103

SEG I Vertical 58 64 62 62 63 SEG II Vertical 70 77 75 75 76 SEG III Vertical 65 72 70 70 71 SEG IV Vertical 24 27 26 26 26

Sub Total 217 240 233 234 236

TOTAL 305 337 328 333 339

D&C spending should increase 10% as well in 2004 to $88 billion. For the reasons listed above, Spears expects spending to plateau through the balance of the period:

D&C Spending by Segment (Millions) 2003 2004 2005 2006 2007

SEG I $11 $12 $12 $12 $12 SEG II $16 $18 $17 $17 $18 SEG III $21 $23 $23 $23 $23 SEG IV $32 $35 $34 $36 $37

TOTAL $80 $88 $86 $88 $90

2 Forecast is subject to change depending on oil and gas prices going forward. This forecast is based on Spears’ June 2004 Drilling and Production Outlook.

Spending on directional & MWD services has grown strongly over the last 7 years and the future looks sound:

Directional +MWD Market LWD $3,000 $900

$800 $2,500 $700

$2,000 $600 $500 $1,500 $400

$1,000 $300 $200 $500 $100 $0 $0

The “health” of the DD+MWD+LWD market dictates the cycle of the downhole drilling tools business, which is the category where we track drilling motors, bottomhole assemblies, MWD systems, fishing tool manufacturing and so on. As the following chart indicates, the downhole drilling tool market peaked in 2001 and has been on a rebound in 2003, a rebound that should extend into 2004 and beyond:

Downhole Drilling Tool Market $600

$500

$400

$300

$200

$100

While we are certain that the oilfield service sector will continue to cycle into the future and that some of the wildest cycling will be seen within the directional drilling segment, we believe that the short, medium and long term futures of the market segment will continue to be positive and that the customer will seek more and more downhole technology solutions to avoid buying expensive drilling rig days. For the directional drilling services and equipment market we project a 6% annual growth through 2007.

Production Technologies

Demand for high-end production technologies is generally driven by these factors:

o Multi-zone completions o Development of complex reservoirs o Deepwater development drilling o High operating cost environments

We have no easily accessible metrics to show that increasingly complex reservoirs are being developed around the world, nor is there a measure for multi-zone completions. We can, however, show how other technologies that are related to these types of developments have grown or been adopted over the years. For example, intelligent completions – sometimes called “smart wells” – have grown from fewer than 5 per year a decade ago to 35- 40 per year currently - an annual growth rate of 22%3:

New "Smart Well" Installations 50

40 WellDynamics Others 30

Systems 10 0

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Another indicator of the move toward more exotic wellbores is the trend in spending on completion equipment and services – packers, multilateral junctions, sliding sleeves and, as shown above, smart completions. Spending on completions is accelerating faster (8% per year) than the growth in drilling activity, suggesting a move toward complex wells:

Completion Equipment & Service $3,500

$3,000

$2,500

$2,000

$1,500

$1,000

$500

3 Source: Spears & Associates, Inc., September 2003. Graph represents annual sales of WellDynamics (Halliburton and Shell joint venture) and all others, including Baker Hughes, Schlumberger, Weatherford and BJ Services. This market may be peaking. WellDynamics now makes more money on parts and accessories than on complete systems.

Production is moving into deeper and deeper waters. Today zones are being tested that lie in waters almost 2 miles deep. As the chart below indicates, the deepwater producing well population is growing about 6% per year:

Deepwater Producing Wells 1000 Brazil 900 Africa 800 GOM 700

600 500

Wells 400

300 200

100 0

The highest cost regions, in terms of drilling and operating costs, are offshore. Offshore drilling has been on a secular growth trend for years and is projected to continue at an average annual growth rate of 4% per year, as seen on the following graph:

Offshore Wells

4,500

4,000 3,500 3,000 2,500 2,000 Wells 1,500 1,000

500

1990 1992 1994 1996 1998 2000 2002 2004 2006

Drilling technology: Active vibration dampener

APPLICATION: Increasing drilling speed is a major economic advantage in all drilling, as is the protection of expensive BHA components from the shock and vibration environment. The best applications for AVD are deep, hard rock holes and high day rate drilling, like deepwater and international, when LWD and RS systems are in the hole.

CUSTOMER: The operating company drilling engineers will mandate the use of these systems as a result of faster drilling. Delivery to the field will be either through the existing directional companies, specialty rental companies, or through an APS-established service group.

VALUE: Increased ROP and the mitigation of shock and vibration are a part of every drilling operation, but some situations are more problematic than others. Drilling engineers recognize the important value of keeping the bit properly engaged at the desired instantaneous weight-on-bit, and in protecting the integrity of their systems. The efficiencies of faster ROPs and less idle time from unnecessary trips create a lower cost well for the customer.

COMPETITION: No system is on the market that actively measures and dampens shock and vibration. Baker Hughes and others have devices that measure these parameters, allowing the driller to modify the drilling process, but no system evaluates and solves the problem downhole. The primary competition is from doing nothing at all.

ADDRESSABLE MARKET: The AVD is designed to work in all Segment III and IV applications where high day rate drilling costs can be significantly reduced through improved drilling efficiencies. AVD also applies to about half the Segment II wells and a small fraction of the low-end wells:

Active Vibration Dampener Addressable Global Drilling Market

High

II IV

Technical Difficulty I III

Low High 15 Drilling & Completion Cost AVD addresses about 60% of the footage drilled each year – almost 200 million of the 305 million feet drilled in 2003:

AVD-addressable market by segment (Millions of Feet) 2003 2004 2005 2006 2007

SEG I Directional 2 2 2 2 2 SEG II Directional 17 19 18 19 20 SEG III Directional 28 31 30 31 32 SEG IV Directional 10 11 11 12 13

Sub Total 57 63 61 64 66

SEG I Vertical 6 6 6 6 6 SEG II Vertical 35 39 38 38 38 SEG III Vertical 65 72 70 70 71 SEG IV Vertical 24 27 26 26 26

Sub Total 130 143 139 140 141

TOTAL 186 206 200 204 207

EXPECTED GROWTH: We believe that the market will rapidly adopt an active vibration dampener once it has been proven to work. As the product matures we expect it to become a familiar component in the package of downhole services found at the rig site.